Mirror visual feedback (MVF) was first proposed as a therapy to relieve amputee phantom limb pain in the early 1990s. It is increasingly used to treat a range of other chronic pain conditions. The evidence base to date is limited. Much of the literature consists of pilot projects or case study designs although larger randomized controlled trails are now emerging. However, the described protocols for MVF are inadequate to adapt to clinical practice. In addition, the therapist sees a heterogeneous population whose characteristics may fall outside those of the tight inclusion/exclusion criteria of research studies. This article provides the theoretical background to MVF and a detailed description of applying this therapy in clinical practice.